Start from the task you need, keep the source and listen before you publish.
Use a scene prompt to describe the result you need. Sign in to save a private project. Add your voice accepts browser-decodable mono or stereo audio up to 25 MB and 120 seconds. Uploaded audio is converted to WAV, then stored in your project for mixing.
The quote identifies speech, singing or remixing and reads the credit price from server configuration. A service that is not configured cannot be confirmed. Changing the script or pronunciation needs a new voice; changing the included bed or its level reuses the dry source.
Open My creations or a recent project to restore saved messages, quotes and versions. A queued task can start once. An ambiguous provider submission is never automatically repeated. Keep its task ID and ask the site operator to review it before requesting another take.
The four audio tools work without sign-in. They do not upload your file. Choose an input, process it, listen and download WAV before refreshing. The normalizer uses RMS with sample-peak protection; it does not certify broadcast LUFS compliance.
This installation may not have a connected conversation, voice or music service. You can still use local audio editing and, with an account, mix your own uploaded recording. No demonstration audio is substituted for unavailable generation.